EP93XX and TS-72XX

EP93XX and TS-72XX

[Christian Gagneraud]

Hi all,

I'm using boards made by TS (Technologic systems), these boards are 
based on the Cirrus Logic EP9301 and EP9302.
TS provide a version of redboot for their boards and of course they 
provide the modified source of ecos (2.0). Part of the path set is 
likely to be more or less what's pending in [1].
I would be interested to see the changes ported to ecos 3.0 and merged 
upstream.

I saw on this ML archive that other peoples are working on EP93XX, and 
that there is some issue with copyright stuff, hopefully it will be 
solved soon.

In the mean time can anyone point me to an unofficial tree, so that i 
can start working on adding the TS specific stuff on top of it?

Regards,
Chris

[1] http://bugs.ecos.sourceware.org/show_bug.cgi?id=1000739

Re: EP93XX and TS-72XX

[John Dallaway]

Hi Christian

Christian Gagneraud wrote:

> I'm using boards made by TS (Technologic systems), these boards are
> based on the Cirrus Logic EP9301 and EP9302.
> TS provide a version of redboot for their boards and of course they
> provide the modified source of ecos (2.0). Part of the path set is
> likely to be more or less what's pending in [1].
> I would be interested to see the changes ported to ecos 3.0 and merged
> upstream.

I suggest you talk with Technologic Systems about this in the first
instance since all code contributions are subject to copyright assignment.

> I saw on this ML archive that other peoples are working on EP93XX, and
> that there is some issue with copyright stuff, hopefully it will be
> solved soon.
> 
> In the mean time can anyone point me to an unofficial tree, so that i
> can start working on adding the TS specific stuff on top of it?

I am not aware of any more recent public sources than those in bug
1000739. As you note, there are several developers already making use of
this work. I would anticipate further contributions as soon as the
copyright assignment is in place and the original contribution is
accepted into eCos CVS.

John Dallaway

Re: EP93XX and TS-72XX

[Christian Gagneraud]

John Dallaway wrote:
> Hi Christian
> 

Hi John,

> Christian Gagneraud wrote:
> 
>> I'm using boards made by TS (Technologic systems), these boards are
>> based on the Cirrus Logic EP9301 and EP9302.
>> TS provide a version of redboot for their boards and of course they
>> provide the modified source of ecos (2.0). Part of the path set is
>> likely to be more or less what's pending in [1].
>> I would be interested to see the changes ported to ecos 3.0 and merged
>> upstream.
> 
> I suggest you talk with Technologic Systems about this in the first
> instance since all code contributions are subject to copyright assignment.

Correct me if I'm wrong, to be integrated upstream, the copyright have 
to be granted to the FSF, is that right? Ecos v2.0 was copyright Red 
Hat and licensed under the eCos License and/or the GPL, and now v3.0 
is copyright FSF and is licensed under the GPL only(?).

By greping through the patch I've generated myself, i've found:

Mixed copyright:
  - Cirrus logic (from bug report 1000739)
  - Red Hat, Inc. (I guess these are files copied over and then adapted)
  - Technologic System (at least redboot/version.c, where they've 
added a copyright print statement)

Authors and contributor:
  - gthomas
  - hmt
  - Jesse Off
  - jordan
  - jskov
  - nickg
  - Travis C. Furrer
  - rgorsegn

So the best would be to ask TS if they can provide officially a patch 
against an official version of ecos with all the copyright statements 
updated.

What do you think?

Regards,
Chris

> 
>> I saw on this ML archive that other peoples are working on EP93XX, and
>> that there is some issue with copyright stuff, hopefully it will be
>> solved soon.
>>
>> In the mean time can anyone point me to an unofficial tree, so that i
>> can start working on adding the TS specific stuff on top of it?
> 
> I am not aware of any more recent public sources than those in bug
> 1000739. As you note, there are several developers already making use of
> this work. I would anticipate further contributions as soon as the
> copyright assignment is in place and the original contribution is
> accepted into eCos CVS.
> 
> John Dallaway

Re: EP93XX and TS-72XX

[John Dallaway]

Hi Christian

Christian Gagneraud wrote:

> John Dallaway wrote:
> 
>> I suggest you talk with Technologic Systems about this in the first
>> instance since all code contributions are subject to copyright
>> assignment.
> 
> Correct me if I'm wrong, to be integrated upstream, the copyright have
> to be granted to the FSF, is that right?

Correct. Copyright in all contributions to the public eCos sources must
be assigned to the FSF as part of the contribution process.

> Ecos v2.0 was copyright Red Hat
> and licensed under the eCos License and/or the GPL, and now v3.0 is
> copyright FSF and is licensed under the GPL only(?).

Not exactly. The run-time sources in eCos 2.0 and later releases are
licensed under a modified version of the GPL with a very important
exception clause. Ref:

  http://ecos.sourceware.org/license-overview.html

> By greping through the patch I've generated myself, i've found:
> 
> Mixed copyright:
>  - Cirrus logic (from bug report 1000739)

The contributor will assign copyright to the FSF as part of the
contribution process. Hopefully no problem here.

>  - Red Hat, Inc. (I guess these are files copied over and then adapted)

Red Hat have assigned all their copyright in eCos to the FSF since eCos
2.0 was released. No problem.

>  - Technologic System (at least redboot/version.c, where they've added a
> copyright print statement)

All changes and additions made by Technologic Systems would need to be
identified (regardless of whether they've added a copyright line) and
assigned to the FSF.

> Authors and contributor:
>  - gthomas
>  - hmt
>  - Jesse Off
>  - jordan
>  - jskov
>  - nickg
>  - Travis C. Furrer
>  - rgorsegn

The above contributors will have already assigned copyright directly or
indirectly to the FSF.

> So the best would be to ask TS if they can provide officially a patch
> against an official version of ecos with all the copyright statements
> updated.

For avoidance of doubt, the current holders of copyright in the platform
port, relevant drivers and any other contributed code (presumably
Technologic Systems) would also need to provide a copyright assignment. Ref:

  http://ecos.sourceware.org/assign.html

A patch against the HEAD of the eCos CVS repository is preferable.

John Dallaway

RE: EP93XX and TS-72XX

[H Hartley Sweeten]

On Monday, September 28, 2009 9:37 AM, John Dallaway wrote:
> Hi Christian
> 
> Christian Gagneraud wrote:
> 
>> John Dallaway wrote:
>> 
>>> I suggest you talk with Technologic Systems about this in the first
>>> instance since all code contributions are subject to copyright
>>> assignment.
>> 
>> Correct me if I'm wrong, to be integrated upstream, the copyright have
>> to be granted to the FSF, is that right?
>
> Correct. Copyright in all contributions to the public eCos sources must
> be assigned to the FSF as part of the contribution process.
>
>> Ecos v2.0 was copyright Red Hat
>> and licensed under the eCos License and/or the GPL, and now v3.0 is
>> copyright FSF and is licensed under the GPL only(?).
>
> Not exactly. The run-time sources in eCos 2.0 and later releases are
> licensed under a modified version of the GPL with a very important
> exception clause. Ref:
>
>   http://ecos.sourceware.org/license-overview.html
>
>> By greping through the patch I've generated myself, i've found:
>> 
>> Mixed copyright:
>>  - Cirrus logic (from bug report 1000739)
>
> The contributor will assign copyright to the FSF as part of the
> contribution process. Hopefully no problem here.
>
>>  - Red Hat, Inc. (I guess these are files copied over and then adapted)
>
> Red Hat have assigned all their copyright in eCos to the FSF since eCos
> 2.0 was released. No problem.
>
>>  - Technologic System (at least redboot/version.c, where they've added a
>> copyright print statement)
>
> All changes and additions made by Technologic Systems would need to be
> identified (regardless of whether they've added a copyright line) and
> assigned to the FSF.
>
>> Authors and contributor:
>>  - gthomas
>>  - hmt
>>  - Jesse Off
>>  - jordan
>>  - jskov
>>  - nickg
>>  - Travis C. Furrer
>>  - rgorsegn
>
> The above contributors will have already assigned copyright directly or
> indirectly to the FSF.
>
>> So the best would be to ask TS if they can provide officially a patch
>> against an official version of ecos with all the copyright statements
>> updated.
>
> For avoidance of doubt, the current holders of copyright in the platform
> port, relevant drivers and any other contributed code (presumably
> Technologic Systems) would also need to provide a copyright assignment. Ref:
>
>   http://ecos.sourceware.org/assign.html
>
> A patch against the HEAD of the eCos CVS repository is preferable.
>
> John Dallaway

The Technologic Systems patch is probably based on the original ep93xx port
from Cirrus Logic.  Brian Austin put the patch into bugzilla on 4/10/2009
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1000739 but the FSF paperwork
never got filed.  Around 9/17/2009 I asked about this again and he said that
the necessary paperwork was going out.  Hopefully the eCos group will be
notified soon and the patch can get merged.  It will need quite a bit of
cleanup but everything "is" there to get RedBoot functioning on the ep93xx.

Regards,
Hartley

Re: EP93XX and TS-72XX

[Christian Gagneraud]

John Dallaway wrote:
> Hi Christian
> 
> Christian Gagneraud wrote:
> 
>> John Dallaway wrote:
>>
>>> I suggest you talk with Technologic Systems about this in the first
>>> instance since all code contributions are subject to copyright
>>> assignment.
>> Correct me if I'm wrong, to be integrated upstream, the copyright have
>> to be granted to the FSF, is that right?
> 
> Correct. Copyright in all contributions to the public eCos sources must
> be assigned to the FSF as part of the contribution process.
> 
>> Ecos v2.0 was copyright Red Hat
>> and licensed under the eCos License and/or the GPL, and now v3.0 is
>> copyright FSF and is licensed under the GPL only(?).
> 
> Not exactly. The run-time sources in eCos 2.0 and later releases are
> licensed under a modified version of the GPL with a very important
> exception clause. Ref:
> 
>   http://ecos.sourceware.org/license-overview.html
> 
>> By greping through the patch I've generated myself, i've found:
>>
>> Mixed copyright:
>>  - Cirrus logic (from bug report 1000739)
> 
> The contributor will assign copyright to the FSF as part of the
> contribution process. Hopefully no problem here.
> 
>>  - Red Hat, Inc. (I guess these are files copied over and then adapted)
> 
> Red Hat have assigned all their copyright in eCos to the FSF since eCos
> 2.0 was released. No problem.
> 
>>  - Technologic System (at least redboot/version.c, where they've added a
>> copyright print statement)
> 
> All changes and additions made by Technologic Systems would need to be
> identified (regardless of whether they've added a copyright line) and
> assigned to the FSF.
> 
>> Authors and contributor:
>>  - gthomas
>>  - hmt
>>  - Jesse Off
>>  - jordan
>>  - jskov
>>  - nickg
>>  - Travis C. Furrer
>>  - rgorsegn
> 
> The above contributors will have already assigned copyright directly or
> indirectly to the FSF.
> 
>> So the best would be to ask TS if they can provide officially a patch
>> against an official version of ecos with all the copyright statements
>> updated.
> 
> For avoidance of doubt, the current holders of copyright in the platform
> port, relevant drivers and any other contributed code (presumably
> Technologic Systems) would also need to provide a copyright assignment. Ref:
> 
>   http://ecos.sourceware.org/assign.html
> 
> A patch against the HEAD of the eCos CVS repository is preferable.

OK. I see 3 different possible scenario:

  - They accept to publish a patch agains CVS HEAD or decent recent 
version (>=3.0) *AND* they assign copyright to the FSF *AND* they 
provide the copyright assignment

Ideal case.

  - Same but for the version of ecos they based their port on.

Still OK, because we can re-use their work and port ourself to ecos CVS.

  - None of them, status-quo
In that case, it means someone would have to port the TS-72XX himself 
once the EP93xx is there or while the EP93XX is ongoing, and *without* 
re-using code from TS own port.

Thanks for the clarifications.
Chris

> 
> John Dallaway